Discussion

MS bearings in TKA have received increasing attention as an evolutionary modification to fixed bearings to better replicate native tibiofemoral kinematics. While the concept has undergone a storied history in its implementation through various products introduced to market from the mid-1990s [12], contemporary systematic reviews have identified comparable midterm performance to standard fixed bearing designs in prosthesis survival and patient-reported outcomes [15]. The evaluation of this second-generation MS device (SAIPH Knee) builds on its first-generation predecessor, the medial rotation knee introduced in 1994, which demonstrates relatively low revision rates at 10-year follow-up (2.56%) [16]. The staged introduction of the present device has verified the medial pivot kinematics of the implant [17] and returned encouraging clinical outcomes in a designer-surgeon series [18]. Subsequent investigations by nondesigner surgeons reported improved clinical knee stability (N = 60) [19] and superior clinical results (N = 46) [20] over contemporary fixed-bearing devices.

All TKA designs have sought to provide stability, longevity, and normal kinematics and the MS knee concept has achieved these aims [13,17,19]. Measures of success are now more aligned to the patient’s perception with goals to reduce the proportion of patients who are unsatisfied with their TKA. By comparing PROMs and satisfaction among patients with MS and non-MS knee designs, Jones et al. [19] concluded knee stability had the greatest influence on patient outcome measures. Pritchett reported that patients can notice the difference and cite stability as a reason for preferring a MS knee over posterior stabilized and cruciate retaining (CR) knees [7]. However, wider literature [[9], [10], [11], [12],15] has shown that inclusion of a medial ball-and-socket articulation in a TKA design does not guarantee improved outcomes. In our study, we evaluated whether the introduction of a second-generation MS knee design to wider use by multiple surgeons could produce consistently high outcomes as reported by the patients. While our study did not assess differences between different MS knee designs and constructs, or the techniques employed to implant them, all surgeons used the same device that had been developed based on experience with the first-generation design [17], with good education prior to use, complete excision of the PCL, and availability only of cemented components. The present analysis represents a definitive multicenter evaluation of this prosthesis in a nondesigner series and the largest cohort study on this category of bearing, comparable to all pooled studies to-date (N = 630) [21]. In line with the hypothesis, the results demonstrated high levels of postoperative function and patient satisfaction, as well as reasonably low joint pain and awareness at 2-year follow-up.

Complications recorded in our study were comparable to literature that attributes most complications to patient factors and pre- and postoperative care [22]. In our study, special attention was given to complications citing stiffness, given the additional constraint in the implant design and surgeons’ limited experience with the concept. In the present series, one case was revised for stiffness, and 2.3% of cases were treated for stiffness by manipulation under anesthetic. Overall, the median ROM at 2 years (119°) was comparable to the average of previous medial-pivot (PCL-sacrificing) designs (117°) [21] and within the range for a broad collection of implant designs at final follow-up [9], as well as the upper boundary of a reference trajectory after TKA [23]. Nevertheless, a proportion of patients (2.3%) reported <100° of maximum flexion at the 2-year follow-up, which is a lower incidence than reported in a localized registry at up to 6 months of follow-up (18.5%) [22]. The MS design is intended to be implanted with more ligament laxity in flexion than contemporary CR designs, without loss of stability [20], and may be less forgiving of a tight flexion gap. Attempts to retain the PCL have been also shown to result in poorer outcomes [19]. Thus, complete and thorough resection of the entire PCL and more physiological ligament laxity, particularly on the lateral side in flexion, are prerequisites for optimal balancing of this TKA, which may translate to a learning curve for surgeons practiced in CR knee designs. We found that using the study device and with appropriate training, increased rates of stiffness were avoided. The relatively few occurrences of stiffness recorded in this study did not represent a measurable learning curve.

Indeed, there was no detectable learning curve within any measure in this study. Although the literature tends to focus on operative time as a convenient indicator of surgeon efficiency during uptake of a new device or technique modification [24], it does not represent the important indicators of patient outcome. There is a risk of increased rate of revision in cases performed as part of the learning curve [25], as well as higher complication rates [26] following the introduction of a new device; however, this was not observed in the present series. The lack of discernible differences in subsamples representing different parts of the surgeon learning curve align with other larger studies examining the learning curve after the introduction of a new implant into practice [27]. In their large cohort study (N = 2000), Whittaker et al. [27] detected clinically insignificant differences for PROMs, including the KOOS Activities of Daily Living subscore and no differences for complications. Given these findings, surgeons taking up the present implant of interest may expect an increase in operative time initially, but a general lack of learning curve with respect to patient-centered outcomes. However, larger samples to adequately power learning curve analyses for complications and revisions are required.

The outcomes in this patient cohort are within the reported population variability for pain, function, joint awareness, as well as revision endpoints. It should be noted that the Australian Orthopaedic Association National Joint Replacement Registry population for the device in question overlaps with the present cohort, as well as representing broader usage beyond the staged introduction. The upper confidence limit (95%) of the revision rate at 2-year follow-up in the present series (2.18%) is higher than the limit at 3 years for the same device (2.0%) in Australia [13] and its predecessor design at 3 years in the UK population (1.27%) [16], but this may be at least partly attributable to the differences in sample sizes. Further, the upper limit in the present series is below the upper confidence limit reported for other medial pivot devices (2.5-3.9%) [13]. Initial concerns that the asymmetric constraint in MS TKA might lead to an increase in tibial loosening were alleviated in midterm radiological and outcomes evaluation of the first-generation implant [1]. Previous reviews of revision and failure modes in medial pivot designs in multiple registries worldwide [10] have also identified implant loosening as a key reason for revision; this is in contrast to the lack of revisions for loosening (femur or tibia) in our series.

Patient-reported knee-localized pain and function, as derived from the OKS and KOOS, were comparable to results for broader TKA populations in national registry reports from New Zealand, Australia, Sweden, the United Kingdom, and the United States. The mean OKS score at 1 year in the present study (40) is in line with the 6-month means reported by the New Zealand Joint Registry (37.8) [28], Australian Joint Replacement Registry (37.6) [29], and 1-year results from the Dutch Arthroplasty Register (39.1) [30]. Similarly, the mean KOOS-Pain subscale (86) at 1 year was comparable to the average for 65-74-year-olds (84.8) observed in a primary knee arthroplasty multicenter registry in the United States (N = 3539) [31] and the Swedish Arthroplasty Register at 1-year (78) [32]. Further, the change in physical activity rating (University of California Los Angeles activity score) of 2 points is in line (1.9 points) with reports for patients >55 years [33] after primary TKA.

Behrend et al. [34] introduced the FJS as a measure of patients’ ability to forget their joint in everyday life, proposing it to be the ultimate goal to ensure maximum patients’ satisfaction. Their TKA patients reported a FJS of 50, whereas healthy control subjects of the same age reported a FJS for the knee of 71.7. With a FJS of 70 at 2 years after surgery, this large multicenter cohort achieved a mean FJS representative of the healthy control subjects reported by Behrend et al. Such scores after TKA have been reported elsewhere [35,36], but it is notable that when compared to their standard practice CR knees, three of the surgeon study participants reported a higher FJS for this MS knee at the 1-year time point [19,20]. This is consistent with findings of a recent meta-analysis reporting a significantly greater FJS for MS knees than non-MS knees [11]. Tso et al. [11] also found the FJS to represent the greatest difference among PROMs scores between knee types. Regarding our observed improvement from 1 to 2 years, Carlson et al. reported similar improvements over the same period but found that the FJS later declined (76.4-64.4) from 2 to 5 years postoperation [36]. With the designer series for this device reporting a high FJS of 75.3 at 5 years [18], follow-up of this cohort to 5 years would be of interest.

While the FJS reflects functional outcome, patient satisfaction after TKA is multifactorial and is influenced by expectations and factors beyond function [37]. Contemporary literature suggests an average dissatisfaction rate after primary TKA of 10%, which remains a high proportion of patients, but may vary in relation to the incidence of known risk factors within a given population [37], such as obesity [38]. The average body mass index observed in this study places the cohort in the obese (class I) category. Nevertheless, good-excellent satisfaction rates at 2-year follow-up (94.5%) compared favorably to a localized Australian registry (N = 2226) at 6-month follow-up (89%) [39], which also identified a potential relationship between osteoarthritis severity and likelihood of dissatisfaction after primary TKA. The present results also compare favorably to multiple studies of satisfaction after TKA (89-96.2%) [14] with these rates expected to be stable over medium to long-term follow-up. The high level of patient satisfaction in our cohort is consistent with the designer series for this MS knee [18].

The findings should be interpreted within the scope and limitations of the study. The findings demonstrate positive outcomes out to 2-year follow-up, and these results are expected to be stable into medium term follow-up [14]. However, further monitoring is necessary to identify late-term issues. The present results should also be viewed within the context of biases that are prevalent in observational studies of this kind. The influence of selection bias on the part of surgeons (patient selection) and the patients themselves (accepting the study and implant) cannot be ruled out entirely; however, the direction and magnitude of its effect on the results cannot be determined with the information available. Surgeon bias was somewhat mitigated by some surgeons using pseudorandomization techniques to select patients for inclusion in the study [20], the consecutive manner of recruitment overall, and the multicenter study design. Some of the analyses (eg, for learning curve) may have been constrained by a lack of power and should be interpreted with some caution. Meaningful comparisons to benchmarks for complications and revision incidence require larger samples to test hypotheses with available methods. Nevertheless, the results demonstrate that this MS knee design can produce consistently good clinical results when used by multiple surgeons in their general practice. The strengths of this study were the large cohort of patients evaluated, and the number of surgeons and hospitals involved. Despite intrinsic differences in prior experience and standard practice among participating surgeons, clinical outcomes for the SAIPH Knee were consistent and comparable with previously published literature for the same device including those for the designer surgeon series [18], supporting its wider adoption.